Modelling nitrogen cycles of farming systems as basis of site- and farm-specific nitrogen management

Abstract

The paper describes a model designed for analysing interrelated nitrogen (N) fluxes in farming systems. It combines the partial N balance, farm gate balance, barn balance and soil surface balance, in order to analyse all relevant N fluxes between the subsystems soil–plant–animal–environment and to reflect conclusive and consistent management systems. Such a system approach allows identifying the causes of varying N surplus and N utilisation.

The REPRO model has been applied in the experimental farm Scheyern in southern Germany, which had been subdivided into an organic (org) and a conventional (con) farming system in 1992. Detailed series of long-term measuring data are available for the experimental farm, which have been used for evaluating the software for its efficiency and applicability under very different management, yet nearly equal site conditions.

The organic farm is multi-structured with a legume-based crop rotation (N₂ fixation: 83 kg ha⁻¹ yr⁻¹). The livestock density is 1.4 LSU ha⁻¹. The farm is oriented on closed mass cycles.

The conventional farm is a simple-structured cash crop system based on mineral N (N input 145 kg ha⁻¹ yr⁻¹). Averaging the years 1999–2002, the organic crop rotation reached, with regard to the harvested products, about 81% (6.9 Mg ha⁻¹ yr⁻¹) of the DM yield and about 93% (140 kg ha⁻¹ yr⁻¹) of the N removal of the conventional rotation. Related to the cropped area, the N surplus calculated for the organic rotation was 38 kg ha⁻¹ yr⁻¹ versus 44 kg ha⁻¹ yr⁻¹ for the conventional rotation. The N utilisation reached 0.77 (org) and 0.79 (con), respectively. The different structure of the farms favoured an enhancement of the soil organic nitrogen stock (35 kg ha⁻¹ yr⁻¹) in the organic crop rotation and caused a decline in the conventional system (−24 kg ha⁻¹ yr⁻¹). Taking account of these changes, which were substantiated by measurements, N surplus in the organic rotation decreased to 3 kg ha⁻¹ yr⁻¹, while it increased to 68 kg ha⁻¹ yr⁻¹ in the conventional system. The adjusted N utilisation value amounted to 0.98 (org) and 0.69 (con), respectively.

Introduction

The intensification of agriculture has led to considerable yield increases, but also induced several environmental problems (Tilman et al., 2001). Many of them like the nitrate load of ground water, emissions of ammonia and greenhouse gases into the atmosphere and the eutrophication of ecosystems are related to high nitrogen (N) inputs in farming systems. Objections have been raised to insufficient N utilisation in agriculture and excessive N emissions from the viewpoint of environmental protection (Isermann, 1990, Van der Ploeg et al., 1997, Crutzen et al., 2008). Numerous measures and mitigation strategies were recommended, however, without adequate success (Isermann, 1994, Eichler and Schulz, 1998).

In crop production, N input is one of the most important yield enhancing factors, but also has an enormous relevance for the environment. Especially in N intensive farming systems the spatial and temporal optimisation of N fertilisation is difficult. Despite the latest application equipment and scientifically based fertilisation algorithms, N utilisation by crop production is incomplete; mainly under unfavourable weather conditions large amounts of the applied N cannot be used by the plants. These N quantities might be stored in the soil organic nitrogen stock (SON) or are subject to gaseous or leaching loss processes. In Germany, the N surplus and N utilisation by soil surface balance are currently 84 kg N ha⁻¹ yr⁻¹ and 48%; at farm gate scale 102 kg N ha⁻¹ yr⁻¹ and 38%, respectively (Osterburg, 2008).

In animal husbandry, N utilisation reaches only 10% in cattle fattening and up to 35% in dairy farming. The excreta contain 65–90% of the N supplied with the feed. Therefore, it is necessary to recycle these N quantities with losses as low as possible. Here, the often practised separation between crop production and animal husbandry is counterproductive. In regions with intensive stock keeping (above 2 LSU ha⁻¹) strained N fluxes with high N losses may occur.

Completely different is the situation in low-input systems like organic farming. Here, nitrogen often is a yield limiting factor. Negative N balances are recorded when the N outputs with cash crops exceed the N inputs from symbiotic N₂ fixation and organic fertiliser, thus reducing the SON stock. Additionally, N cycles in organic farming systems may be intensified by the use of biogas plants or intensive animal husbandry.

The examples show the great necessity of optimising the N cycles in farming systems. In the past, legal restrictions could not sufficiently reduce N emissions. Further success will only be possible by active participation of the farmers. Rising fertiliser prices will orient the farmers’ economic interest towards an improvement of N efficiency. But also the growing awareness of the environment favours a critical rethinking of farm-specific N management.

This leads to the question of which management tools are suited for optimising N cycles. The most frequently used method for analysing farming systems is N balancing (Schröder et al., 2003). The calculated N surplus levels reflect the loss potential of reactive N compounds approximately (Oenema et al., 2003). In simple input-output accounting systems farms are regarded as “black box”, where only inputs and outputs at the farm gate are quantified. These tools neglect internal farm structures, nutrient fluxes and management practices. In order to support farm management decisions, it is essential to describe the farm internal structures and the relationships between soil–plant–animal–environment (Küstermann et al., 2008). Thus, it becomes possible to disclose the causes of different N utilisation, to identify weak spots and to elaborate strategies for optimising N cycles on-farm scale (Oenema et al., 2003, Watson et al., 2002).

Yet, the numerous attempts of N balancing have so far not yielded a standardised method, since the objectives are very diverse. As a consequence, results are only comparable if the balancing methods have been clearly defined. According to Halberg et al. (2005) and Goodlass et al. (2003), N balancing models can be classified according to:

• System level and spatial breakdown: farm scale (farm gate balance), crop production (field balance or soil surface balance), livestock keeping (barn balance).

• N fluxes and N pools considered.

• Balancing coefficients and algorithms.

• The database: measured values, estimates, assumptions or statistical data.

The paper deals with the introduction and application of a N balancing model designed for N management at the farm level. It describes agricultural farms as systems which respond to interferences like structural changes, alterations in intensity and technology. All subsystems of a farm (soil–plant–animal–environment) are linked via N fluxes which allow, for example, the simulation of interactions between crop production and animal husbandry.

To verify the efficiency and validity of the model, it has been used in the experimental farm Scheyern in southern Germany. Here, a long-term experiment was started in 1992 comprising an organic (org) and a conventional (con) farming system, with the aim of analysing their impacts on ecosystems (Schröder et al., 2002, Schröder et al., 2008). During the experimental period different use intensities (phases of intensification and extensification) were applied. All management information has been recorded in detail; the resulting effects on N cycles were analysed by measurements and modelling of all relevant N fluxes and N pools.

The model allows one to estimate soil surface balance, barn balance and farm gate balance and to link these partial balances to an aggregated system balance, in order to evaluate management effects on the N surplus, N utilisation and N emissions and to show optimisation potentials. Measurement records help to estimate the degree of accuracy achieved with the model in describing N cycles and in disclosing the extent of errors in N balance sheets. General conclusions are drawn on the recommended procedure of N balancing, in order to make optimal use of them in farm management.